A buyer standing outside a listing pulls out a phone, taps the number on the sign, and lands on the agent's website. Not on office wifi. Not on a laptop. On whatever cellular signal that block happens to get, on a screen the width of a hand.
That moment is the one that actually decides whether the site works, and it is almost never the moment an agent tests.
Google only sees the mobile version now
Most agents assume Google looks at their site the way a visitor with a good connection does: full page, fast load, everything rendering cleanly. That assumption is wrong in a specific, checkable way.
Under mobile-first indexing, Google's own crawler uses the mobile version of a page to decide how it indexes and ranks that page, everywhere, including in desktop search results. The desktop version is not a separate track that gets its own evaluation. If the mobile version is missing content, hiding a section, or loading slowly, that becomes the baseline Google works from for the whole site.
This matters more for real estate than for most other kinds of small business sites, because agent websites lean heavily on visual content: listing photo galleries, neighbourhood maps, IDX search widgets. All of that has to survive the mobile version, on a mobile connection, or Google is evaluating a weaker page than the one shown to a desktop visitor at a listing presentation.
The desktop test hides the real conditions
Ask most agents how their site performs and they will describe checking it on their office computer. That test measures almost nothing about what a buyer actually experiences.
Office wifi is fast, stable, and close to the router. A phone standing outside a house, or in a car between showings, is on a cellular connection that varies by carrier, by block, and by how many other devices are competing for the same tower. The same page that loads in a blink on a desk can take several times longer on a weak signal, and nothing about a desktop test reveals that gap.
The honest test is different and takes two minutes: turn wifi off on your phone, load your homepage and one listing page using cellular data only, away from your building. Time it. That number, not the score from a desktop speed tool, is closer to what a buyer standing at a curb actually sees.
The buyer's situation shortens the patience, too
Speed problems on mobile compound with something that has nothing to do with bandwidth: the visitor's situation.
Someone browsing from a couch on a Saturday morning has time and a comfortable position. Someone standing outside a house they are curious about, phone in one hand, is mid-errand. They are more likely to be interrupted, more likely to have a specific short-lived intent (does this agent handle this neighbourhood, is there a number I can tap right now), and more likely to abandon a slow page for the next search result rather than wait it out.
That difference in patience is why the same three-second delay costs more on a phone in the field than on a desktop at home. The technical fix is the same either way, but the business cost of skipping it is higher on the visit that is actually happening most often.
The 70% figure, and why the trend line matters more than the number
NAR's 2025 Profile of Home Buyers and Sellers puts mobile and tablet use during the home search at 70%. On its own, a single-year figure is a snapshot. The trend behind it is the more useful part: that share moved from 57% in 2020 to 70% in 2025, climbing every year measured in between.
A site tuned around desktop assumptions was already behind buyer behaviour a few years ago. It is further behind now, and the direction of the trend means waiting does not close the gap on its own.
Plenty of buyers still research from a laptop in the evening, and a listing presentation is often viewed on a desktop screen in front of the seller, so desktop still matters. The mobile version can no longer be treated as the secondary case that gets checked last, because for most agents it is now the majority case.
What actually breaks on cellular data that does not break on wifi
Total file size is the variable that behaves differently between the two connection types. Wifi tolerates a heavy page reasonably well because bandwidth is high and stable. Cellular data does not offer that same margin, and the gap between a fast connection and a weak one is far wider on cellular than on wifi.
A few specific things that suffer most:
A full listing photo gallery, especially one where images were uploaded straight from a camera without resizing. On wifi it loads slowly but tolerably. On a weak cellular signal, several of the later images may never finish loading before the visitor gives up.
An embedded walkthrough video that autoplays. It competes for the same limited bandwidth as everything else on the page and is often the single heaviest element on a listing page.
A third-party IDX search widget loaded on every page rather than only where it is used. Each instance re-downloads its own scripts, and that repeated cost is far more noticeable on a weak connection than on a fast one.
None of these are exotic problems. They are common, and each one is worse specifically on the connection type most visitors are actually using, which is the reason mobile data deserves its own check rather than being folded into a general speed audit.
Testing with the right settings matters as much as the right device
A common mistake undoes the value of testing on a phone at all: testing at home, where the phone quietly connects to fast wifi and behaves like a desktop machine.
If you use a browser-based speed testing tool instead of a physical phone, most default to a desktop connection profile unless changed. Switching the test to a mobile device profile with a throttled connection, a setting available in most free testing tools, is often the difference between a passing score and a page that reveals real problems. Run the test both ways once, so you know exactly what gap you are working with.
The field test remains the more honest one. A tool gives you a number. Standing outside a house with your own phone on cellular data, waiting for your own contact form to appear, tells you what a buyer actually experiences, and it costs nothing.
What this changes about loading priority
Treating mobile as the primary version, rather than the adapted one, changes what should load first on a page and how fast the whole page finishes.
The phone number and a way to make contact should be visible without scrolling and should load early, before the rest of the photo gallery. The map or address should not depend on a heavy third-party embed finishing first. Anything decorative, a background video, an animated section, a large hero image that exists mostly for atmosphere, should be allowed to load last, because a buyer standing at a curb needs the contact information before the atmosphere.
That is a design decision as much as a technical one, and it is easy to get backwards when a site was built by looking at a desktop mockup first and treating the phone version as a smaller copy of the same layout. Our post on what buyers actually check on an agent's website covers the content side of that same priority question.
Where speed connects to conversion
A site that loads acceptably on mobile data still has to do something with the visitor once it arrives, and that is a separate problem from speed.
If the number of clicks it takes to reach a phone number or a form on a mobile screen is high, or the form itself asks for more than a name and a way to reach back, the speed work does not matter, because the visitor never reaches the point of contact. Our post on real estate website forms that convert covers how many fields a mobile visitor will actually complete, and lead response time covers what happens after they do.
A fast site with no clear next step and a slow site that loads before anyone sees it are different failures with the same outcome: a visit that produces nothing. Speed removes one obstacle. It does not replace the others.
The takeaway
Google evaluates the mobile version of a real estate website, not the desktop one, and most buyers are now searching on that same mobile version, on a cellular connection that behaves differently from office wifi. Testing on a desk with good wifi hides both of those facts at once. Test on your own phone, with wifi off, standing somewhere your signal is average rather than excellent, and treat what you see there as the real state of your site. That is the version a buyer meets, and it is the only one that counts for both ranking and the call.
If your current site was built around a desktop mockup, a free rebuild preview will show what changes when mobile is the starting point instead of an afterthought. For questions about the rebuild process, see the FAQ or get in touch.
